U.S. patent number 7,448,542 [Application Number 12/115,238] was granted by the patent office on 2008-11-11 for method for detecting a non-scan at a retail checkout station.
This patent grant is currently assigned to International Business Machines Corporation. Invention is credited to Russell P. Bobbitt, Myron Flickner, Arun Hampapur, Charles A. Otto, Sharathchandra U. Pankanti, Unsang Park, Akira Yanagawa, Yun Zhai.
United States Patent |
7,448,542 |
Bobbitt , et al. |
November 11, 2008 |
Method for detecting a non-scan at a retail checkout station
Abstract
A method for detecting a non-scan at a retail checkout station
includes detecting the passing of an item across a scanner device
of a retail checkout station, obtaining an image of the item
passing across the scanner, detecting a scan of an item passing
across the scanner to establish a scanned item, and establishing a
register associated with the scanned item. A scan occurs if the
image of the item passing across the scanner substantially matches
the register associated with the scanned item. Conversely, a
non-scan is triggered when the image of the item passing across the
scanner does not match the register associated with the scanned
item.
Inventors: |
Bobbitt; Russell P.
(Pleasantville, NY), Flickner; Myron (San Jose, CA),
Hampapur; Arun (Norwalk, CT), Otto; Charles A. (Lansing,
MI), Pankanti; Sharathchandra U. (Norwalk, CT), Park;
Unsang (East Lansing, MI), Yanagawa; Akira (New York,
NY), Zhai; Yun (White Plains, NY) |
Assignee: |
International Business Machines
Corporation (Armonk, NY)
|
Family
ID: |
39940719 |
Appl.
No.: |
12/115,238 |
Filed: |
May 5, 2008 |
Current U.S.
Class: |
235/383; 186/61;
235/454; 235/462.01; 348/150; 705/23 |
Current CPC
Class: |
G06Q
20/208 (20130101); G07G 1/0054 (20130101); G07G
3/006 (20130101) |
Current International
Class: |
G06K
15/00 (20060101); A63F 9/02 (20060101); G06K
7/10 (20060101); G06K 7/14 (20060101); G06Q
20/00 (20060101); H04N 7/18 (20060101); H04N
9/47 (20060101) |
Field of
Search: |
;235/462.11,462.31,462.41,462.14 ;705/18,23 ;348/150
;186/59,61 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
Primary Examiner: Fureman; Jared J
Attorney, Agent or Firm: Cantor Colburn LLP Wardas; Mark
Claims
The invention claimed is:
1. A method for detecting a non-scan at a retail checkout station
comprising: detecting the passing of an item across a scanner
device of a retail checkout station; determining that the item
passing across the scanner device was not registered as a scan;
establishing a potential non-scanned item based on the item not
being registered as a scan; obtaining an image of the potential
non-scanned item; establishing a scanned item based on an item
passing across the scanner device being registered as a scan;
extracting features associated with the scanned item; comparing the
features associated with the scanned item with the image of the
potential non-scanned item; and triggering an actual non-scan if
the features of the scanned item do not substantially match the
image of the potential non-scanned item.
2. The method of claim 1, further comprising: detecting an object
not associated with the potential non-scanned item; and removing
the object not associated with the potential non-scanned from the
image of the potential non-scanned item.
3. The method of claim 2, wherein detecting the object not
associated with the potential non-scanned item includes detecting a
portion of a cashier moving the potential non-scanned item across
the scanner device.
4. The method of claim 1, further comprising: detecting multiple
attempts to scan an item; establishing multiple potential
non-scanned items based on the item not being registered as a scan;
and evaluating each of the multiple potential non-scanned items
prior to indicating an actual non-scan of the item.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to the art of retail checkout stations and,
more particularly, to a method of detecting a non-scan at a retail
checkout station.
2. Description of Background
It is well known that shrinkage is a significant problem in the
retail industry. It has been observed that a large proportion of
retail shrinkage is a result of cashier fraud. Current methods to
detect cashier fraud depend on human supervision. Human supervision
is both costly and non-scalable as the number of cashiers
increases. Even without considering labor and scalability issues,
some types of cashier fraud, such as "non-scans" are difficult for
humans to detect.
A non-scan occurs when a cashier intentionally fails to scan, often
referred to as "sweet hearting" or unintentionally, i.e.,
accidentally, fails to properly pass a product over a barcode
scanner region of the checkout station. In general, there are two
methods by which intentional non-scans are committed. The first
method involves positioning a product so that a product barcode is
pointed away from the scanner. An alternative method is to cover
the barcode so that the scanner is unable to detect it. To the
casual observer, a non-scan appears to be a legitimate scan.
Unintentional non-scans typically result from a lack of attention
on the part of the cashier.
A human might perhaps detect intentional and unintentional
non-scans by being extremely vigilant or through the aid of a video
surveillance system with cameras positioned over the barcode
scanners. In either case, detection requires careful observation of
the cashier, either directly or via captured images. Thus, a single
individual would only be able to monitor one cashier or camera at a
time to detect fraud. In order for such a system to be
cost-effective, on average each cashier would have to contribute to
retail shrinkage at a rate greater than the pay rate of a loss
prevention (LP) officer assigned to observe the cashier. An
additional issue exists in the amount of tedium involved in these
observations. It is unlikely that a human observer would be capable
of maintaining high attention levels for extended periods.
In addition to visual monitoring, auditory monitoring is another
potential method of preventing loss. Barcode scanners often emit an
audible beep when a proper scan occurs. Thus, an LP officer could
listen for the presence of a beep for each item being scanned. In
such a case, the LP officer would need to be in close proximity to
the cashier being observed, i.e., each cashier would require a
dedicated LP officer. A dedicated LP officer at each checkout
station would add considerably to resource allocation and in no way
mitigate the tedium discussed above. Furthermore, it is often
difficult or impossible to differentiate between sounds when
multiple cashiers are in close proximity, or if the ambient
environment is noisy.
SUMMARY OF THE INVENTION
The shortcomings of the prior art are overcome and additional
advantages are provided through the provision of a method for
detecting a non-scan at a retail checkout station. The method
includes detecting the passing of an item across a scanner device
of a retail checkout station, obtaining an image of the item
passing across the scanner, detecting a scan of an item passing
across the scanner to establish a scanned item, and establishing a
feature associated with the scanned item. Establishing that the
image of the item passing across the scanner substantially matches
the feature associated with the scanned item. Triggering a non-scan
if the image of the item passing across the scanner does not match
the feature associated with the scanned item.
Additional features and advantages are realized through the
techniques of exemplary embodiments of the present invention. Other
embodiments and aspects of the invention are described in detail
herein and are considered a part of the claimed invention. For a
better understanding of the invention with advantages and features
thereof, refer to the description and to the drawings.
TECHNICAL EFFECTS
As a result of the summarized invention, technically we have
achieved a solution, which enables a retail storeowner to monitor
cashier activity to reduce shrinkage associated with sweet hearting
or the intentional failure of a cashier to properly scan an item to
be purchased. Non-scan events are evaluated real time or
forensically to determine if a particular cashier(s) is responsible
for inventory shrinkage. In addition, by virtually eliminating
false-positives thereby increasing non-scan detection
reliability.
DESCRIPTION OF THE DRAWINGS
The subject matter which is regarded as the invention is
particularly pointed out and distinctly claimed in the claims at
the conclusion of the specification. The foregoing and other
objects, features, and advantages of the invention are apparent
from the following detailed description taken in conjunction with
the accompanying drawings in which:
FIG. 1 is a right perspective view of a retail checkout station
including a system for detecting a non-scan of an item to be
purchased in accordance with an exemplary embodiment of the present
invention;
FIG. 2 is a flow diagram illustrating a method for detecting a
non-scan of an item to be purchased in accordance with an exemplary
embodiment of the present invention; and
FIG. 3 is a flow diagram illustrating a method of evaluating a
potential non-scan of an item to be purchased in accordance with an
exemplary embodiment of the present invention.
The detailed description explains the exemplary embodiments of the
invention, together with advantages and features, by way of example
with reference to the drawings.
DETAILED DESCRIPTION OF THE INVENTION
With initial reference to FIG. 1, a checkout station constructed in
accordance with an exemplary embodiment of the present invention is
generally shown at 2. Checkout station 2 includes a main housing 4
that establishes a cashier portion 6, a product scanning portion 8,
a product transport portion 12, and a bagging portion 14. Cashier
portion 6 includes a register 20 having a display portion 22, a
printer portion 24 and a cash drawer 25. Cashier portion 6 also
includes an input zone 28 having a keypad/card reader 30, and a
display 33.
Product scanning portion 8 includes a scanning surface 40 having a
product scanner device 41 provided with a plurality of trip wires
44-47 that establish a first region of interest (ROI) 49. Trip
wires 44-47 can be a plurality of energy (light, sound etc) beams
or simple a boundary designating a particular area or reference
zone(s) in an image. Product scanner device 41 scans or reads
product identification codes, such as Universal Product Code (UPC),
industrial symbol(s), alphanumeric character(s) or other indicia
associated with items to be purchased.
Product transport portion 12 includes a conveyor belt 55 that moves
purchased products from product scanner device 41 to bagging
portion 14. Towards that end, conveyor belt 55 includes a first end
portion 57 positioned adjacent product scanner 41 that extends to a
second end portion 60 through an intermediate portion 62. As will
be discussed more fully below, first end portion 57 includes a
plurality of trip wires 66-69 that establish a second ROI 70. Trip
wires 66-69, in a manner similar to that described above, can be
plurality of energy beams or simply a boundary designating a
particular area or reference zone in an image. In accordance with
an exemplary embodiment of the invention, checkout station 2 also
includes a non-scan detect system 72 having a scan detection module
73 that is employed in connection with a camera 74 to detect
non-scans at checkout station 2. At this point it should be
understood that the term "scan" means a successful reading of a
product identification code, e.g., UPC symbol, by a scanner. The
term "non-scan" means an unsuccessful reading of the product
identification code either accidentally, or with intent. In any
event, scan detection module 73 can include a direct connection
between product scanner device 41 and non-scan detect system 72, a
microphone that scans for the typical beep associated with a proper
scan, or a link to a data-base that is updated real-time each time
a bar code or the like is registered at product scanner device
41.
Reference will now be made to FIG. 2 in describing a method for
detecting a non-scan at checkout station 2 employed by non-scan
detection system 72. Initially, camera 74 captures a reference
image of product scanner device 41 with no objects present as
indicated in block 200. After obtaining the reference image, a skin
model is implemented in block 202. As will be discussed more fully
below, the skin model is employed by non-scan detect system 72 to
remove portions of the cashier, i.e., hands, arm, etc., that might
otherwise interfere with captured images. At this point, non-scan
detect system 72 begins scanning for objects passing through first
and second ROI's 49 and 70. Once an object is detected in ROI 49 or
ROI 70 by, for example, detecting the object passing though trip
wires 44-47 and/or 66-69, an image(s) of the object passing through
ROI 49 and/or ROI 70 is/are captured by camera 74 as indicated in
block 204. At this point, a determination is made whether the
object is actually in one of ROI 49 and ROI 70 as indicated on
block 206. An object is sensed within ROI 49 and/or ROI 70 by a
breaking of trip wires 44-47 and/or 66-69. In addition, an object
can be detected by a spatiotemporal relationship of the object in
ROI 49 and/or ROI 70. When the object is present within ROI 49
and/or ROI 70 an optimal image or frame is obtained in block 208.
Once the object is determined to no longer be in one of ROI 49
and/or ROI 70 non-scan detect system 72 determines whether the
object was in proximity to product scanner device 41 in block 210.
If not, scanning continues in block 204, however, if an object was
in proximity to product scanner device 41, non-scan detect system
72 determines whether a proper scan took place.
Often times, an item will pass though ROI 49 and/or ROI 70 multiple
times before a scan is recorded. That is, product scanner device 41
might be dirty, bar codes might be wrinkled, damaged or otherwise
difficult to scan or the cashier might have the object at an
improper angle for scanning. In any event, if the product is not
scanned, e.g., a bar code associated with the product is not
recorded/registered by product scanner 41, as determined in block
220, each non-scan is saved in memory and the optimal frame
captured in block 208 for each non-scan is filtered to remove
objects not associated with the scanned item such as, portions of
the cashier as detected by the skin model in block 222. At this
point, non-scan detect system 72 determines whether a candidate or
potential non-scan exists in block 224. A potential non-scan is
indicated when the optimal frame, as captured in block 208, for
each non-scan is determined to contain a product. Once it has been
determined that an object is present in a given frame, the
optimality of that frame is determined in block 230. The optimality
of the frame is determined based on a variety of techniques, such
as but not limited to, comparing object size to size thresholds or
analyzing object characteristics/features or consistency with
characteristics that are exhibited by products or classes of
products. Product characteristics may be input manually or learned
automatically. The characteristics/features of the object could
also be employed to further enhance the determination that an
object is present within ROI 49 and/or ROI 70. In any case, if no
object is present in the optimal image, non-scan detect system 72
returns to block 204. If, however, a potential non-scan is
indicated in block 224, and an object is detected in block 230,
features of the product, such as color, shape, position, texture,
etc., are extracted from the optimal frame and stored in an
internal data structure (not shown) such as a stack, in
anticipation of the next barcode scanned in block 200. The next bar
code scanned in block 200 establishes a scanned item. At this
point, non-scan detect system 72 evaluates any potential non-scans
against the scanned item in block 234.
Reference will now be made to FIG. 3 in describing a method of
evaluating potential non-scans employed in block 234. At first, a
determination is made whether any potential non-scans of the
product exist in a stack or memory in block 300. If the product was
scanned properly the first time, an a cashier did not intentionally
or unintentionally pass any products over scanning surface 40
without an intervening signal from checkout station 2 indicating a
successful scan of the product, no non-scan exists and evaluation
is complete. If, however, one or more potential non-scans do exist,
evaluation continues. In block 302, each potential non-scan is
extracted and a determination is made whether the product in the
optimal frame matches the feature of the product obtained in block
230. If a match is found, a non-scan is contra-indicated, and the
non-scan is deleted as indicated in block 306, and non-scan detect
system 72 checks for additional non-scans of the product in block
302. If however, a match is not found, an actual non-scan is
indicated in block 320. At this point, non-scan detect system 72
determines whether any additional potential non-scans of the
product exist in memory. If so, each additional potential non-scan
of the product is designated as an actual non-scan in block 322. In
this manner, non-scan detect system virtually eliminates any
occurrence of false-positive indications of a non-scan event.
Once evaluation is complete in block 234, non-scan detect system 72
checks whether there are additional items to process for the
transaction in block 240. If so, scanning and monitoring continue
in block 224. If not, the transaction is complete and a
determination is made whether additional transactions will occur in
checkout station 2 in block 246. If so, non-scan detect system 72
confirms the identity of the cashier in block 248. If a new cashier
is present, a new skin model is set in block 202, else scanning and
monitoring continues. If, in block 246, the transaction is
determined to be the last transaction of a particular period,
scanning and monitoring end.
At this point it should be appreciated that exemplary embodiments
of the present invention provide a system and method for
determining whether non-scans of products are occurring at a
checkout station. The system and method are employed by store
personnel as real time, or forensic indicators of "sweet hearting"
or a cashier giving away merchandise without receiving payment. The
system can be operated with little or no human interaction thereby
increasing detections without the need for a dedicated Loss
Prevention LP officer for each checkout station and/or constant
rotation of LP officers to different roles to limit inattention due
to boredom.
The capabilities of the present invention can be implemented in
software, firmware, hardware or some combination thereof. As one
example, one or more aspects of the present invention can be
included in an article of manufacture (e.g., one or more computer
program products) having, for instance, computer usable media. The
media has embodied therein, for instance, computer readable program
code means for providing and facilitating the capabilities of the
present invention. The article of manufacture can be included as a
part of a computer system or sold separately.
The flow diagrams depicted herein are just examples. There may be
many variations to these diagrams or the steps (or operations)
described therein without departing from the spirit of the
invention. For instance, the steps may be performed in a differing
order, or steps may be added, deleted or modified. All of these
variations are considered a part of the claimed invention.
While the preferred embodiment to the invention has been described,
it will be understood that those skilled in the art, both now and
in the future, may make various improvements and enhancements which
fall within the scope of the claims which follow. These claims
should be construed to maintain the proper protection for the
invention first described.
* * * * *